<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
nmt-mpst-id-en-lr_1e-05-ep_20-seq_128_bs-32
This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.7787
- Bleu: 0.0338
- Meteor: 0.1312
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Meteor |
|---|---|---|---|---|---|
| No log | 1.0 | 202 | 3.1965 | 0.0132 | 0.0696 |
| No log | 2.0 | 404 | 3.0644 | 0.0224 | 0.0975 |
| 3.5509 | 3.0 | 606 | 2.9995 | 0.0255 | 0.1075 |
| 3.5509 | 4.0 | 808 | 2.9538 | 0.0269 | 0.1106 |
| 3.2374 | 5.0 | 1010 | 2.9221 | 0.0277 | 0.1134 |
| 3.2374 | 6.0 | 1212 | 2.8996 | 0.0286 | 0.1165 |
| 3.2374 | 7.0 | 1414 | 2.8750 | 0.0291 | 0.1177 |
| 3.143 | 8.0 | 1616 | 2.8611 | 0.0297 | 0.1197 |
| 3.143 | 9.0 | 1818 | 2.8466 | 0.0303 | 0.1209 |
| 3.092 | 10.0 | 2020 | 2.8330 | 0.0312 | 0.1229 |
| 3.092 | 11.0 | 2222 | 2.8234 | 0.0318 | 0.1247 |
| 3.092 | 12.0 | 2424 | 2.8130 | 0.0322 | 0.1264 |
| 3.0511 | 13.0 | 2626 | 2.8058 | 0.0323 | 0.1269 |
| 3.0511 | 14.0 | 2828 | 2.7970 | 0.0324 | 0.1272 |
| 3.0288 | 15.0 | 3030 | 2.7914 | 0.033 | 0.1288 |
| 3.0288 | 16.0 | 3232 | 2.7877 | 0.0331 | 0.1299 |
| 3.0288 | 17.0 | 3434 | 2.7837 | 0.0333 | 0.1302 |
| 3.0133 | 18.0 | 3636 | 2.7809 | 0.0336 | 0.1308 |
| 3.0133 | 19.0 | 3838 | 2.7792 | 0.0337 | 0.131 |
| 3.0028 | 20.0 | 4040 | 2.7787 | 0.0338 | 0.1312 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Datasets 2.7.0
- Tokenizers 0.13.2
